Single channel and dual channel noise detection systems and techniques

Abstract

An ECG processing system for identifying noisy ECG data segments is provided. The ECG processing system includes a network interface and a processor. The network interface is configured to receive a first collection of ECG data segments from wearable medical devices associated with a plurality of patients, each wearable medical device of the wearable medical devices being configured to be continuously worn by a patient for an extended period of time. The processor is configured to produce a second collection of ECG data segments from the first collection of ECG data segments that excludes noisy ECG data segments.

What is claimed is: 
     
       1. An electrocardiogram (ECG) processing system for processing noisy ECG data segments, comprising:
 one or more network interfaces configured to receive a first collection of ECG data segments from wearable medical devices associated with a plurality of patients, each wearable medical device of the wearable medical devices being configured to be continuously worn by a patient for an extended period of time; 
 a memory configured to store the first collection of ECG data segments; and 
 one or more processors and associated circuitry coupled with the memory, the one or more processors being configured to
 (a) retrieve the first collection of ECG data segments from the memory, 
 (b) for each ECG data segment of the first collection of ECG data segments,
 (i) transform the ECG data segment into a corresponding baseline representation of the ECG data segment by dividing the ECG data segment into a plurality of sample periods, each sample period of the plurality of sample periods spanning between 2 and 8 seconds, and
 for each sample period within the plurality of sample periods, removing ECG data collected within the sample period that transgresses one or more threshold values to generate the baseline representation of the ECG data segment, 
 
 (ii) fit the baseline representation of the ECG data segment to a function comprising at least one first-degree coefficient, 
 (iii) compare the at least one first-degree coefficient of the function to a baseline drift threshold, the at least one first-degree coefficient characterizing baseline drift in the baseline representation of the ECG data segment, and 
 (iv) identify the ECG data segment as a noisy ECG data segment where the at least one first-degree coefficient of the function transgresses the baseline drift threshold, 
 
 (c) produce a second collection of ECG data segments from the first collection of ECG data segments that excludes the noisy ECG data segments, and 
 (d) output one or more ECG metrics relating to the plurality of patients based on the second collection of ECG data segments. 
 
 
     
     
       2. The ECG processing system of  claim 1 , wherein the first collection of ECG data segments is updated over time to include additional ECG data segments. 
     
     
       3. The ECG processing system of  claim 2 , wherein the one or more processors is configured to periodically repeat steps (a)-(d) after a duration of time. 
     
     
       4. The ECG processing system of  claim 3 , wherein duration of time comprises at least one of 30 minutes, 1 hour, 4 hours, 24 hours, 1 week, 2 weeks, and 1 month. 
     
     
       5. The ECG processing system of  claim 3 , wherein the one or more processors is configured to determine one or more trends in the one or more ECG metrics over the duration of time. 
     
     
       6. The ECG processing system of  claim 1 , wherein the function comprises a polynomial. 
     
     
       7. The ECG processing system of  claim 6 , wherein the at least one first-degree coefficient comprises a coefficient of a first-degree monomial of a third-degree polynomial. 
     
     
       8. The ECG processing system of  claim 1 , wherein the baseline drift threshold comprises a range of threshold values between −0.5 and 0.5. 
     
     
       9. The ECG processing system of  claim 1 , wherein the one or more processors are configured to use linear regression to fit the baseline representation of the ECG data segment to the function. 
     
     
       10. The ECG processing system of  claim 1 , wherein each of the ECG data segments spans between 15 and 120 seconds. 
     
     
       11. The ECG processing system of  claim 1 , wherein each of the ECG data segments comprises an average length of at least one of about 15 seconds, about 30 seconds, about 45 seconds, about 60 seconds, about 90 seconds, about 120 seconds, about 150 seconds, about 180 seconds, about 5 minutes, about 10 minutes, and about 15 minutes. 
     
     
       12. The ECG processing system of  claim 1 , wherein the one or more ECG metrics comprises a heart rate metric, a heart rate variability metric, a QRS duration metric, a QT interval metric, a heart rate turbulence metric, and/or pre-ventricular contraction (PVC) burden metric. 
     
     
       13. The ECG processing system of  claim 1 , wherein the one or more processors are configured to provide the one or more ECG metrics to a cardiac prediction process. 
     
     
       14. The ECG processing system of  claim 1 , wherein the one or more ECG metrics can be used by the one or more processors to determine or one or more arrhythmia events comprising ventricular tachycardia, ventricular fibrillation, bradycardia, tachycardia, asystole, pause, bigeminy, and/or trigeminy. 
     
     
       15. The ECG processing system of  claim 1 , further comprising the wearable medical devices, wherein the wearable medical devices comprise dry ECG electrodes configured to acquire ECG signals. 
     
     
       16. The ECG processing system of  claim 1 , further comprising the wearable medical devices, wherein the wearable medical devices comprise adhesively attached ECG electrodes configured to acquire ECG signals. 
     
     
       17. The ECG processing system of  claim 1 , wherein the one or more ECG metrics are analyzed by the one or more processors to determine a current heart failure condition of at least one of the plurality of patients. 
     
     
       18. The ECG processing system of  claim 17 , wherein the one or more ECG metrics are analyzed by the one or more processors to determine a change in the current heart failure condition of the at least one of the plurality of patients. 
     
     
       19. The ECG processing system of  claim 1 , wherein the one or more processors are further configured to:
 determine, for each sample period within the plurality of sample periods, a mean of ECG data collected within the sample period; 
 determine, for each sample period within the plurality of sample periods, a standard deviation of ECG data collected within the sample period; and 
 determine, for each sample period within the plurality of sample periods, the baseline drift threshold based on the mean of the ECG data collected within the sample period and the standard deviation of the ECG data collected within the sample period. 
 
     
     
       20. The ECG processing system of  claim 1 , further comprising the wearable medical devices, wherein the wearable medical devices comprise a first wearable medical device associated with a first heart failure patient and a second wearable medical device associated with a second heart failure patient.